Previous modeling studies revealed that the hydrometeorological cycle will be enhanced in a warm climate, with marked variability in weather conditions. Despite its importance in the regional water cycle, the progress in warm-season quantitative precipitation forecasts (QPFs) has been slow due to the dominant weak-gradient large-scale forcing but strong thermodynamic conditions. The objectives of this study are to a) demonstrate the predictability of weekly to monthly variations of clouds and precipitation during June 1998 in the LSA-East using an improved, high-resolution, coupled MM5/SSiB model, and b) examine the spatial-temporal structures of the simulated clouds/ precipitation in relation to the large-scale forcing, fine orography and heterogeneous surface conditions. The LSA-East in June 1998 was characterized by the passage of seven cyclones and the development of three mesoscale convective systems (MCSs), causing the regional widespread anomalous precipitation with local flooding conditions. To achieve a realistic simulation of the regional climate over the LSA-East, we found it is necessary to modify the SSiB scheme as follows: a) the negative moisture in the prognostic equation of specific humidity is adjusted to ensure mass conservation; b) the dissipative heating at the bottom boundary is included; c) the surface resistance law is parameterized following Zhang and Anthes (1982); d) the unpaired vegetation and soil types are used; and e) vegetation cover fraction is allowed to vary based on observations. A series of sensitivity experiments using coarse resolutions showed that the above modifications of the model physics reduce significantly the errors in the rainfall and surface temperature fields. It is found that the revised version of the coupled MM5/SSiB model could reproduce quite realistically the climate conditions over the LSA-East. In particular, the model captures well the passage of seven cyclones and the development of three mesoscale convective systems (MCSs), as observed. The temporally-averaged precipitation and maximum/minimum surface air temperatures are well simulated, as verified against the observed. We have also compared the model-simulated sea-level pressure field at the daily and weekly time scale to the observed, and found that the model reproduces reasonably well the passage of various perturbations. Examination of the upper-level potential vorticity (PV) field indicates that some precipitation events were well correlated to the PV anomalies. The vertical bias profiles, averaged over the LSA-East, indicate that the temperature and moisture biases using the SSiB scheme are significantly smaller than the simulations using the other schemes. Some case studies, selected from the monthly simulations, reveal the interaction of topography, internal gravity waves, and surface moisture anomalies with cloud microphysics in the generation of precipitation.